AI

The Saplings AI MCP Recipe turns your Drupal site into a fully compliant Model Context Protocol (MCP) server by installing and configuring AI Core, AI Agents, the MCP module, and other required dependencies. Once applied, your Drupal installation can talk directly to MCP-aware clients, execute AI functions, and invoke agents, extending your site with a wide range of AI-powered workflows. Drupal’s robust content model and modular architecture make it an ideal MCP server, especially when its content is exposed as resources and combined with AI tools for dynamic content generation and management.

Features

• Installs AI Core, AI Agents, the MCP module, and other necessary modules.
• Configures a general-purpose MCP server to support seamless interaction with connected MCP clients.
• Enables execution of AI functions and invocation of agents from your Drupal site.
• Allows customization so Drupal content can be exposed as MCP resources and used as AI tools for workflow automation.
• Keeps configuration open-ended so you can further tune the MCP module, such as pointing it at different LLM providers or adding custom tools and resources as your use cases grow.

Installation Instructions

1. Preparation
Ensure your Drupal installation is up to date. It is recommended to take a full backup of your site before applying the recipe.

2. Dependencies
Verify that Composer is installed on your server, since several dependencies are managed through it.

3. Install the recipe

Use Composer to install the Saplings AI MCP recipe:

composer require drupal/ai_saplings_mcp

4. Enable the modules

After installation, enable the recipe and its modules using Drush (or through the Drupal admin UI):

drush recipe ../recipes/ai_saplings_mcp

5. Test the configuration

Confirm that your setup is working by connecting an MCP client to your Drupal site and testing basic interactions such as listing resources or invoking a simple tool.

Further Configuration

The recipe provides a sensible default MCP server configuration, but you are encouraged to refine it to match your environment:

• Adjust the MCP server settings provided by the MCP module, including authentication and endpoint options.
• Swap in different LLM backends or AI providers through Drupal’s AI configuration if you prefer alternate models.
• Define additional tools and resources that expose Drupal content, search, or custom workflows to your MCP clients.

Once configured, your Drupal site becomes a powerful, AI-aware MCP server that can plug cleanly into modern LLM tooling.

SiteDash is a central dashboard for developers, agencies, and site owners managing multiple Drupal and WordPress sites. It's currently in development. Join the interest list at SiteDash.ai.

The Problem with Multiple Sites

Managing more than a handful of sites means juggling hosting dashboards, repository interfaces, third-party services, and spreadsheets that are never quite up to date. Someone asks which sites need security updates, and you're opening ten browser tabs.

SiteDash aggregates site data in one interface. It connects to the Audit Export Module for Drupal and the Audit Export plugin for WordPress, pulling configuration details, module versions, update status, and hosting information into a unified view. The module supports Drupal 7 through 11, and the plugin works with WordPress 6 and later.

Who This Is For

If you maintain multiple sites, SiteDash gives you clarity. Freelancers with a dozen client projects, agencies balancing portfolios of fifty sites, or product teams needing governance across their web properties all benefit from centralized visibility.

Unified Site Data

Connect your sites and see metadata, configurations, and integrations from one dashboard. No more hunting through wikis or ticket systems for basic information about what's installed where.

SiteDash integrates with code hosts like GitHub and Bitbucket, hosting platforms like Acquia and Pantheon, CI tools like CircleCI and Tugboat, and project management systems including Jira, ClickUp, and Teamwork. It also connects to AI providers like Anthropic and OpenAI, Google Workspace tools, analytics platforms, and operations services like UptimeRobot.

LLM-Powered Infrastructure Queries

SiteDash includes an MCP server that lets AI tools access all your indexed site data through a single connection. Instead of wiring each service independently, your LLM-backed assistants can query everything at once.

Ask "Which sites are running Drupal 10.2.x?" or "Show me all sites with the Webform module enabled" and get immediate answers across your entire portfolio. The MCP integration enables AI-assisted onboarding, configuration audits, and infrastructure planning without manually checking each site.

Self-Service Access for Non-Technical Users

Once a developer connects sites to SiteDash, project managers and non-technical stakeholders can query site information on their own. Instead of filing tickets to ask which modules are installed or which sites need attention, they use natural language queries through the MCP-connected AI interface. The setup work happens once, then anyone on the team can get answers without waiting for developer availability.

Streamlined Agency Onboarding

Agencies often track dozens of client sites in spreadsheets. When a new project arrives, someone spends hours documenting what's installed: modules, themes, hosting details, third-party integrations, and potential technical debt.

The Audit Export Module and plugin automate that snapshot. SiteDash indexes the data and keeps it current. Paired with the MCP server, an LLM can handle much of the discovery work, querying indexed data and surfacing what matters without digging through admin screens.

What Changes

You spend less time context-switching between tools and more time making decisions. Teams share a unified source of truth instead of confirming basic facts in Slack. New developers get situational awareness without hunting through scattered documentation. AI-assisted querying means faster audits and planning.

Early Access

SiteDash is actively evolving. Join the interest list at SiteDash.ai to get notified as development progresses. The goal is an open alpha testing phase starting in January 2026.